<![CDATA[@thomasfuchs @emilymbender This article is about a study of the use of #LLM #AI in radiology and other medical areas.]]>@thomasfuchs @emilymbender
This article is about a study of the use of in radiology and other medical areas. The study’s conclusions confirm what we already know: LLMs do not (cannot) reason. They sound plausible, but they will say anything so long as it seems plausible. They do not “know” the conceptual meaning of anything, including truth and falsehood. They might make a plausible “explanation” of evidence and deductive reasoning, but they are fundamentally unable to DO what they SAY. They are not built that way.

is a scientific field that prides itself on data, hypothesis, reasoning, and evidence for conclusions. This rapid adoption of AI in medicine should be anathema to doctors.

Check the terms and conditions of your admission papers if you have to go to a hospital (as I did, for a minor emergency). Deep in the fine print you will find they disclose using AI for transcription and “other documentation”…by third parties…offsite. JFC! Pardon me. It is so buried they are clearly hoping it does not become a point of contention. But I insist, and you will too after you read this. https://hachyderm.io/@thomasfuchs/116494546658408106

]]>https://postcall.pub/topic/d2a46e76-b7c7-40a1-bc7e-7452fb0e1e37/@thomasfuchs-@emilymbender-this-article-is-about-a-study-of-the-use-of-llm-ai-in-radiology-and-other-medical-areas.RSS for NodeFri, 01 May 2026 12:10:39 GMTThu, 30 Apr 2026 22:20:29 GMT60